Embedded systems have been around for nearly 40 years and, over that time, there has always been the need for embedded software development tools. In the beginning, the source of these tools was almost exclusively the chip manufacturers. Companies like Intel, Motorola and Zilog supplied the development software and all the necessary kit to do debugging – anyone remember the ubiquitous “blue boxes”? In those days, most programming was in assembly language, which was necessarily chip-specific.
But things changed and, in the 1980s, a number of companies started offering complete development systems for a variety of chips – notably Hewlett Packard [later Agilent] and Tektronix. These systems included software tools, a dedicated computer to run them on and an in-circuit emulator [ICE] for target debugging. Gradually, these companies, who had hardware development backgrounds, began to back off from the software parts of their products …
The result was the appearance of a number of embedded software development tools companies, who only sold software. An early example was Microtec Research, whom I joined in 1986; the company was acquired by Mentor Graphics 10 years later, which is how I come to be where I am today. Interestingly, at the same time, a number of smaller ICE manufacturers also appeared – examples include MICE/Microtek, Zax and Applied Microsystems [AMC].
As time went by, through the 1990s and into the new century, software gradually became an increasingly dominant part of an embedded system design and the software companies flourished. At the same time, changes in technology rendered in-circuit emulation mostly obsolete and some other, erstwhile familiar names disappeared. In the last few years there has been another trend: many chip manufacturers want to get back into the software business. Interestingly, although this is a viable business in its own right, their main motivation is to drive sales of silicon [IMHO].
Is this trend beneficial to the end customer – the developers of embedded systems? I think not. By being somewhat locked in to a given software environment, they are unwittingly locked into specific silicon too, which limits their long term opportunities for innovation. But I believe these users are too smart to be caught out by such a lock in. This view is borne out by a recent internal market study.
For some years, Mentor Graphics has endeavored to increase visibility in the embedded software market – the branding Mentor Embedded for the Embedded Software Division was driven by that motivation. This strategy is bearing fruit. The study shows that, market awareness of Mentor Embedded has steadily risen, putting us in second place, after Wind River. This makes us the strongest brand among chip-independent embedded software companies, as Wind River is owned by Intel. I believe that this ownership will be their downfall and I look forward to the next study in a year or so to see whether I am proved right.